CSV extension

The csv extension enables reading of CSV files.

This extension is included by default in the CLI, Python, and WebAssembly clients for GlareDB. A csv schema will be created automatically containing all CSV related functions.

Reading CSV files

The read_csv table function can be used to read a CSV file:

SELECT * FROM read_csv('path/to/file.csv');

The read_csv file can also be used to read TSV (Tab-Separated Values) files as well:

SELECT * FROM read_csv('path/to/file.tsv');

When reading a CSV or TSV file, the "dialect" used will be inferred. This includes inferring the delimiter (comma for CSV, tab for TSV) as well as if the file contains a header. Once the dialect has been determined, the types for each record will then be inferred.

DESCRIBE can be used to determine the types that were inferred for each column:

DESCRIBE read_csv('path/to/file.csv');

read_csv is an alias for the namespaced csv.read function, and can be used interchangeably:

SELECT * FROM csv.read('path/to/file.csv');

If your CSV files ends with .csv or .tsv, the function call can be omitted. The function to use will be inferred automatically:

SELECT * FROM 'path/to/file.csv';

Multiple files can be provided using either a list of files, or a glob. All files are currently expected to have the same schema.

To read a specific set of files:

SELECT * FROM read_csv(['file1.csv', 'file2.csv']);

To read all CSV files in data/:

SELECT * FROM read_csv('data/*.csv');

If the glob ends with .csv or .tsv, the function call can be omitted:

SELECT * FROM 'data/*.csv';

Function reference

All CSV table functions are located in the csv schema. A complete list can be found using list_functions:

SELECT *
FROM list_functions()
WHERE schema_name = 'csv';

csv.read

Alias: read_csv, csv_scan, csv.scan.

The read_csv function takes a path to a CSV file and returns a table containing the parsed data.

SELECT * FROM read_csv('userdata1.csv');

By default, read_csv will:

• Automatically infer column data types
• Detect and skip the header row (if present)

Additional parameters can be provided for other file systems. For example, we can provide AWS credentials for accessing a CSV file in a private S3 bucket:

SELECT * FROM csv.read('s3://bucket-name/path/to/file.csv',
                       region='us-east-1',
                       access_key_id='YOUR_ACCESS_KEY',
                       secret_access_key='YOUR_SECRET_KEY');

Note: Support for custom parsing options, including explicit type definitions and header row overrides, will be available soon.